We report the results of UI-EWD as a tumor bleeding therapy in 41 patients over a three-year period. In our cohort, UI-EWD was applied as a salvage therapy for tumor bleeding when conventional hemostasis was not sufficient or as a monotherapy in cases unsuitable for conventional hemostatic modalities. UI-EWD was found to have an excellent immediate hemostatic success rate (97.5%) and a low re-bleeding rate (7.5% and 22.5% on days 7 and 28, respectively). To the best of our knowledge, the present study is one of the largest to be conducted on the use of a hemostatic powder to treat GI tumor bleeding.
Tumor bleeding presents challenges in clinical settings because conventional endoscopic hemostasis is not always effective . According to limited reports on the subject, the initial hemostasis success rate of conventional endoscopic therapy for tumor bleeding ranges from 31 ~ 40% and the short-term re-bleeding rate is about 80% [16, 17]. Hemostasis failure of conventional endoscopic modalities is associated with tumor size, friability, pathological angiogenesis, or tumor induced necrosis [18,19,20]. In addition, the acidic stomach environment and pancreatic juice can promote tumor bleeding because they dissolve blood clots and digest tumor tissues, which lack a protective barrier epithelium and mucous . When endoscopic hemostasis fails, more invasive treatments, such as surgery or interventional angiography are required for hemostasis [20, 21]. However, surgery has a high mortality rate  and interventional angiography has a poor clinical success rate  and introduces the risk of complications such as ischemic organ damage . Accordingly, there is need for a noble, effective endoscopic hemostatic modality for tumor bleeding. In the present study, UI-EWD was found to have an excellent immediate hemostasis success rate, which is in accord with previous findings [10, 24]. This high success rate can be explained by the hydrogel nature of UI-EWD. Usually, hemostasis is difficult to achieve at tumor bleeding sites by conventional hemostasis because bleeding surfaces are large with multiple bleeding points [25, 26]. The hydrogel formed by UI-EWD covers entire tumor bleeding surfaces and prevents contact with acidic media, pancreatic juice, and bile.
The relatively high re-bleeding rate (33–49%) [8, 10, 11] of tumors after immediate hemostasis presents a technical challenge to previously existing hemostatic powders. The 30-day re-bleeding rate in our study was 22.4%, which is lower than the re-bleeding rates of other commercially available hemostatic powders (47%) in high risk, non-variceal UGIB patients . Furthermore, the 6-month cumulative survival rate obtained in our study was 73.2% (n = 30) and the follow up period was adequate (Table 1). We attribute this lower re-bleeding rate to strong tissue adhesion exhibited by UI-EWD, which remains well attached despite intestinal peristalsis. UI-EWD hydrogel has been reported to be present at 70.2% of sprayed bleeding sites by second-look endoscopy at 24 h . Therefore, we believe UI-EWD provides much better attachment than other available hemostatic powders, which makes it a useful tool for the treatment of tumor bleeding and promising option in terms of reducing re-bleeding rates. In a retrospective case series on malignant GI bleeding, 30-day mortality rates of 14% to 48% were reported [18, 28, 29]. Re-bleeding is known to be associated with poor survival in patients with hemostasis [15, 30]. In the current study, two patients experienced re-bleeding and received additional UI-EWD treatment up to 3 times as bridge therapy. These patients eventually underwent chemotherapy and surgery and survived for more than a year. Therefore, we assume that UI-EWD can be considered a bridge therapy modality in selective patients even if re-bleeding occurs after immediate hemostasis of tumor bleeding has been achieved.
Because the purpose of endoscopic therapy in patients with tumor bleeding is salvage or bridge therapy, complications caused by the treatment modality during tumor bleeding control may adversely affect prognosis. In a previous study, perforation was reported to be a severe adverse event of other commercially available hemostatic powders , and there are also concerns about the risks of obstruction and systemic embolization . Furthermore, perforations resulting from the use of other commercial hemostatic powders has been attributed to tissue rupture and perforation caused by high spraying pressures . UI-EWD was specially designed to eliminate the effect high application pressures. No perforation or tissue damage associated with hemostatic powder application was observed in our cohort, and no other hemostatic powder related complications were observed.
Of the 41 patients included in this study, 18 patients (43.9%) were treated with UI-EWD as a salvage therapy (Fig. 1), and electrocoagulation was used in more than half of these patients. UI-EWD was applied as a monotherapy in 23 (56.1%) patients and immediate hemostasis success rates were similar in those treated with UI-EWD as a salvage therapy. Re-bleeding rates in these two groups at 7 and 28 days were 4.3% (n = 1) and 26.1% (n = 6), which was not inauspicious. On the other hand, Forrest Ib was the most frequent bleeding type in the monotherapy group. There was no case of Forrest type Ia. To more comprehensively assess the efficacy of UI-EWD monotherapy as a salvage therapy for tumor bleeding for different Forrest types a well-designed, large-scale prospective study is needed.
This current study has some limitations. First, the study is inherently limited by its retrospective design and lack of a control cohort. However, there are already many reported researches on conventional endoscopic hemostasis for tumor bleeding, this study shown the results that can be considered sufficiently effective. Second, the number of study subjects was relatively small. However, previous studies on the effects of hemostatic powders on tumor bleeding enrolled fewer than 20 patients . In fact, the present study was conducted on a larger cohort and with longer follow up than previous studies on the topic. Third, tumor bleeding locations varied, but despite this heterogeneity, immediate hemostasis was successful in all by one patient, in who initial bleeding control failed. Furthermore, this result also suggests UI-EWD is applicable at any location in cases of upper GI tumor bleeding, and that its hemostatic effects are independent of location. Forth, we couldn’t figure out the possible reasons of re-bleeding after successful immediate hemostasis. There were only 9 patients in the re-bleeding group, and there was no significant difference in patients and bleeding characteristics between the group with or without re-bleeding. (Additional file 1: Table 1 and 2) A large-scale, well-designed randomized controlled study is needed to determine the possible factors of re-bleeding after immediate hemostasis.